Re: st: pre and post event coding

[Nick Cox <njcoxstata@gmail.com>]

Stata uses the terminology of variables, not columns, and
observations, not rows. We are not in spreadsheet land any more!

With your variable names, try this:

tsset GVKEY FYEAR
gen d_c_prev = L.cashta if downgrade
gen d_c_next = F.cashta if downgrade
gen u_c_prev = L.cashta if upgrade
gen u_c_next = F.cashta if upgrade

and check out

-help tsset-

-help varlist-

and

http://www.stata.com/support/faqs/data-management/true-and-false/

Nick
njcoxstata@gmail.com


On 12 December 2013 21:28, Khieu, Hinh <hdkhieu@usi.edu> wrote:
> Dear Fernando, Nick, and other statalist members,
>
> Thanks for your responses. I am pasting 40 observations below. I need a column for lag of cashta and a column of cashta one year after around downgrade. I also need 2 more columns of cashta before and after upgrade. A total of 4 columns. Could you just help me with one year lead and one year lag? I will figure out more years of lags and leads. For example, the value of second row of cashta is 0.0534525 and should be the in the lag cashta column for upgrades since it is the cash balance one year before upgrade =1 for firm gvkey (an ID in the database) 002812. Similarly, the value of the 4th row of cashta is 0.011070578 and should be in the lead cashta column for upgrades since it is one year after upgrade=1 for the same firm. The two lead and lag cashta just created should be in the same row or something so that I will create a third column for myself which is lead minus lag. I hope this is clearer and greatly appreciate any help with stata coding.
>
> Thanks,
> Hinh
>
> GVKEY   FYEAR   cashta  upgrade downgrade
> 002812  1987    0.033511614     0       0
> 002812  1988    0.0534525       0       0
> 002812  1989    0.019296537     1       0
> 002812  1990    0.011070578     0       0
> 002812  1991    0.014896329     0       0
> 002812  1992    0.018504308     1       0
> 002812  1993    0.011067856     0       0
> 002812  1994    0.012099207     0       0
> 002812  1995    0.029543541     0       0
> 002812  1996    0.013809676     0       0
> 002812  1997    0.012663689     0       0
> 002812  1998    0.012127395     0       0
> 002812  1999    0.013981739     0       0
> 002812  2000    0.01000064      0       0
> 002812  2001    0.131549467     0       0
> 002812  2002    0.213038699     0       0
> 002812  2003    0.008395931     0       1
> 002812  2004    0.01828809      0       0
> 008596  1996    0.069620253     0       0
> 008596  1997    0.060511541     0       0
> 008596  1998    0.050087362     0       0
> 010884  1987    0.158100174     0       0
> 010884  1988    0.134588069     1       0
> 010884  1989    0.179119977     0       0
> 010884  1990    0.160353232     0       0
> 010884  1991    0.264202242     0       0
> 010884  1992    0.143434548     1       0
> 010884  1993    0.055176807     0       0
> 010884  1994    0.061631187     0       1
> 010884  1995    0.10345515      0       0
> 010884  1996    0.115754212     0       1
> 010884  1997    0.052340656     0       0
> 010884  1998    0.035432956     0       0
> 010884  1999    0.037695768     0       0
> 010884  2000    0.040104436     0       0
> 030443  1994    0.03267254      0       0
> 030443  1995    0.016850206     0       1
> 030443  1996    0.036283762     0       1
> 030443  1997    0.107138038     0       0
> 030443  1998    0.03172147      1       0
> 030443  1999    0.025653675     0       0
> 030443  2000    0.008677145     0       0
> 030443  2001    0.010861392     0       0
> 030443  2002    0.007522965     0       0
> 030443  2003    0.034200644     1       0
>
> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Fernando Furquim
> Sent: Thursday, December 12, 2013 10:27 AM
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: pre and post event coding
>
> Hihn,
>
> Could you share some code or data? Remember to -xtset- your data (so that Stata treats it as panel). Then you could use the lag and lead operators (L4 and F4 to lag/lead 4 time periods, for example). Those will lag and lead appropriately accounting for the use of panel data as long as you -xtset- first.
>
> Assuming you have a regular old panel dataset, this should be pretty simple, but if your dataset is mixing time periods or something it'd be trickier.
>
> Good luck!
>
> Fernando
>
> On Wed, Dec 11, 2013 at 10:06 PM, Khieu, Hinh <hdkhieu@usi.edu> wrote:
>> Dear Statalist members,
>>
>> I have a panel data of firm years. The events are credit ratings downgrades and upgrades. I want to identify the cash holdings one year before downgrade (or upgrade) and one year after downgrades (or upgrades). One objective is to construct cash holdings after minus cash holdings before the event. Another objective is to calculate cash holdings volatility 4 years before and cash holdings volatility 4 years after an upgrade or downgrade.
>>
>> I know how to do that in SAS, but do not know how to code it in Stata. I try the [_n-1] and [_n+1] for cash holdings, but it is not necessarily correct because the code does not wrap around an event. Here my downgrades column has 1's for downgrades and 0's otherwise. My upgrades column has 1's for upgrades and 0's otherwise.
>>
>> I'd appreciate any amount of help.
>>
>> Thanks,
>> Hinh
>>
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/help.cgi?search
>> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>> *   http://www.ats.ucla.edu/stat/stata/
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/